Midazolam is a commonly used sedative in critically ill, mechanically ventilated patients in intensive care unit (ICU) settings worldwide. We used a nine-step decision-making algorithm to determine whether therapeutic monitoring of midazolam in the ICU is warranted. Midazolam has a higher clearance and shorter half-life than other benzodiazepines, and prolonged sedation is achieved with continuous infusion. There appears to be very good correlation between plasma concentrations of both midazolam and its active metabolite, alpha1-hydroxymidazolam, and the degree of sedation. However, due to high interpatient variability, it is not possible to predict the level of sedation in any given individual based on plasma concentration of midazolam or its metabolites. Moreover, no simple and practical assay is available to quantitate midazolam plasma concentrations in the acute ICU setting. Many scales are available to assess the sedative effects of midazolam. Because the plasma concentration of midazolam required to achieve a constant level of sedation is highly variable, it is usually more prudent for the clinician to monitor for sedation with a validated clinical scale than by plasma concentrations alone. Various physiologic parameters, including age-related effects, compromised renal function, and liver dysfunction affect the pharmacokinetics of midazolam and alpha1-hydroxymidazolam. Although routine drug monitoring for all critically ill patients receiving midazolam is not recommended, this practice is likely beneficial in patients with neurologic damage in whom sedation cannot be assessed and in patients who have renal failure with a prolonged time to awakening.